Risk Assessment for Hydrogen

1

• Risk Assessment for Hydrogen
• Codes and Standards
• Roger Cox, Jay Keller, Chris Moen
• Sandia National Laboratories
• Jim Ohi
• National Renewable Energy Laboratory
• International Conference on Hydrogen Safety
• Pisa, Italy
• September, 2005

2

• How does the existing codes and standards
  development process incorporate risk in
  determining requirements to provide a given level
  of safety?
• How can risk assessment inform this process in
  setting requirements, e.g., separation distances?
  
• What risk assessment methods and techniques are
  most useful for this process?
• What are the benefits and costs of developing
  risk-informed codes and standards?

3
Workshops at SNL and NREL

• Define safety scenarios and themes that drive
  CS development
• Set priorities for RD and analysis

• Engage stakeholders to explore risk assessment
  tools
• Define requirements for RA in development of
  CS for hydrogen

4
(No Transcript)
5
Stakeholder Participants
6
Risk Accounting
Expert panels
Failure Modes and Effects Analysis - FMEA
Probabilistic Risk Assessment - PRA
7
(No Transcript)
8
Observations

• Probabilistic Risk Assessment (PRA) can
• clearly define what you do not know
• be used to determine appropriate scenarios for
  defining clearance distances and size of
  hazardous zones
• Expected magnitude of risk should drive risk
  assessment from FMEA towards PRA
• PRA needs event frequency data, data, data
• lack of publicly available data on event
  frequency
• Code Development community moves ahead with best
  information available
• from traditional method Panel of Experts to PRA
• concern over acceptability of RA in code
  development process
• cost and schedule (process is schedule-driven but
  moving to using more data)
• RA may be new concept to community

9
Observations

• Risk Standards -- find maximum tolerable limits
• use existing service stations (gasoline, natural
  gas )
• need to make H2 implementation safer than current
  stations
• risk standards not as openly embraced in the US
  as in other countries
• Need generic models of hydrogen systems for risk
  assessment
• Industry performs detailed risk assessments and
  makes risk based decisions
• ensure quality and minimize risk and help put
  products into market
• not explicitly used in codes and standards
  development
• fundamental driver is same as CS development
  safety and liability
• What does it cost to continue code development as
  is? Until costs are understood, cannot know size
  of problem
• what is benefit versus cost of incorporating
  risk-informed code development process as opposed
  to business as usual?

10
Model Code Developers Perspectives

• NFPA
• prioritized research activities of interest to
  NFPA code
• activities stationary power setbacks, metal
  hydride
• storage, system risk analysis, and vehicle
  refueling setbacks
• need to identify the 10 of failure scenarios
  that
• generate 90 of the hazardsneed more on loss
  histories
• Technical Committee asserted that hydrogen poses
  no
• greater risk than does natural gas and set the
  separation
• distances to be the same
• ICC
• ad hoc committee commissioned in 2000 to
  introduce
• new code language for commercial use of
  hydrogen
• ad hoc committee is panel of experts, but they
  want
• information that is easy to understand to
  persuade voting
• body of ICC

11

• Perform high level risk assessment to identify
  unknowns
• safety of industrial technologies and practices
  well-established
• focus on step out technologies where little
  past experience
• public/refueling interfaces
• fuel delivery and PRDs
• emergency response modes
• CDOs expressed interest but needs driven by
  schedule and will consider quantitative
  information, if available, to make decisions
• concern that PRA activities will be hard to
  bring to closure
• CDOs are more interested in qualitative studies
• SDOs may make use of quantitative studies
• CDOs do not need rigorous quantitative
  information to defend code
• language outside of committees

12
Next Steps from Workshop

• Commissioned three working groups of experts to
• define requirements for a high level risk
  assessment to guide priorities
• hydrogen infrastructure system definition
• define maximum tolerable limits of risk (i.e.,
  equivalent to or less than that implicit for
  gasoline and CNG fueling)
• assess cost/benefit of risk-informed CS vs.
  business-as-usual
• identify and compile relevant data, including
  accident event frequencies
• DOE can serve as central data acquisition point
  to encourage submission of data and protect
  business sensitivity of data
• interact directly with standards and code
  committees to inform CS development process with
  RA approaches
• Report Workshop results and plan RA activities
  with Codes and Standards Tech Team

13
Action Items from Workshop

• Define requirements for a high level risk
  assessment to guide priorities
• define maximum tolerable limits of risk (i.e.,
  equivalent to or less than that implicit for
  gasoline and CNG fueling)
• define baseline hydrogen fueling architecture
• Identify and compile relevant data, including
  accident event frequencies
• establish central data acquisition point to
  encourage submission of data and protect business
  sensitivity of data
• Interact with standards and code committees to
  incorporate RA approaches in CS development
  process
• select key standards to assess/modify with RA
  approach
• assess C/B of risk-informed CS vs BAU

14

• How can we efficiently incorporate RA and RD
• information into the CS development process?
  
• how will this information be used to modify
  codes and standards?
• Current RD tasks are focusing on quantifying
• hazards for release events
• where/who/how will information be gathered or
  generated on event frequency?
• What risk thresholds should be used?
• develop a new risk standard for hydrogen?
• adopt risk standard for petroleum refueling?

Requirements needed by someone who would
perform the risk assessment
15

• Identify CS development requirements
• Define problem
• define system
• identify release events
• identify hazards
• define assets to be protected
• define risk management methodology
• identify risk criteria and stakeholder
  requirements
• Quantify hazards
• identify existing data and models
• perform experiments to characterize the unknown
• develop quantitative models and identify
  uncertainties
• Assess risk
• identify event frequencies and quantify
  consequences
• estimate overall system risk
• characterize uncertainty in overall risk
  estimate
• identify system elements that most contribute to
  risk
• identify data most valuable in reducing risk
  uncertainty

16

• Delivery
• liquid spill during delivery
• truck crash or rail derailment
• Bulk Storage
• tank failure large, unconfined releases
• tank breach due to exceeded lifetime
• Dispensing
• electrostatic build-up, discharge
• hose failure
• heat build up and PRD release during refueling
• fueling systems check valve failure
• leak during nozzle disconnect
• fire in adjacent mini-market
• Servicing
• fire during vehicle repair

events rated high priority at 12/03 workshop
17
Work Plan Approach
Probabilistic Risk Assessment - PRA
Data, Haz-Ops, FMEA
Interaction with SDO Expert Panels, other
stakeholders
18
Work Plan Plan Timetable
FY06 1 2 3 4
FY05 1 2 3 4
FY07 1 2 3 4
I. Probabilistic Risk Assessment Gasoline
fueling RA baseline H2 fueling architectures Initi
ating events Consequences Risk vs
architecture Draft PRA (fueling station)
II. Data, Haz-Ops, FMEA Component failure
frequencies Baseline station footprint Iterate
with station template Feedback to PRA
III. Interaction with SDOs/CDOs Reverse RA of
standards RA case studies for new standards C/B
analysis of RA vs BAU Incorporate RA in SDO
process
19
Thank you!
For more information, contact Pat Davis, US
Department of Energy patrick.davis_at_ee.doe.gov C
hris Moen, Sandia National Laboratories chris.moe
n_at_snl.gov Jim Ohi, National Renewable Energy
Laboratory jim_ohi_at_nrel.gov